Qwen3.5-4B-NSFW-Literotica is a specialized large language model fine-tuned for high-fidelity erotic storytelling and creative prose. Based on the Qwen 2.5 architecture (4B parameters), this model has been optimized using Parameter-Efficient Fine-Tuning (PEFT) on a curated corpus of literary erotica.
Training Details
The model was fine-tuned on a high-speed compute cluster using a surgical resume strategy from precise checkpoints to ensure mathematical continuity and optimal convergence.
Target Modules: All linear layers (q, k, v, o, gate, up, down)
Context Length: 512 tokens
Optimization: Paged AdamW (32-bit) with a learning rate of 2e-5.
Usage & Inference
Prompt Format
This model uses the standard ChatML format:
Inference Settings
For optimal creative output, the following parameters are recommended:
Temperature: 0.75
Top-p: 0.90
Repetition Penalty: 1.1
Max New Tokens: 512+
Responsible Use & Disclaimer
This model is intended for adult audiences (18+) and generates explicit NSFW content. By using this model, you acknowledge that you are of legal age in your jurisdiction and that you will use it in compliance with all local laws and regulations. The model is a tool for creative fiction and should not be used to generate illegal or non-consensual content.
